Google launched its much-anticipated new flagship AI mannequin, Gemini 2.0 Pro Experimental, on Wednesday. The announcement was a part of a sequence of different AI mannequin releases. The firm can be making its “reasoning’ mannequin, Gemini 2.0 Flash Thinking, obtainable within the Gemini app.
Notably, Google is releasing these AI fashions because the tech world stays fixated on cheaper AI reasoning fashions supplied by the Chinese AI startup DeepSeek. DeepSeek’s fashions match or surpass the efficiency of main AI fashions supplied by American tech corporations. At the identical time, companies can entry DeepSeek’s fashions by way of the corporate’s API for a relative steal.
Google and DeepSeek each launched AI reasoning fashions in December, however DeepSeek’s R1 received much more consideration. Now, Google could also be making an attempt to place its Gemini 2.0 Flash Thinking mannequin in entrance of extra eyes by way of its common Gemini app.
As for Gemini 2.0 Pro, the successor to the Gemini 1.5 Pro mannequin Google launched final February, Google says that it’s now the main mannequin in its Gemini AI mannequin household.
Google by chance introduced the Gemini 2.0 Pro mannequin’s launch within the Gemini app’s changelog roughly per week in the past. But this time, it’s for actual. The firm is releasing an experimental model of the mannequin on Wednesday in its AI growth platforms, Vertex AI and Google AI Studio. Gemini 2.0 Pro will even be obtainable to subscribers to Gemini Advanced within the Gemini app.
Specifically, the brand new Gemini Pro mannequin excels at coding and dealing with complicated prompts, per Google, and it comes with “higher understanding and reasoning of world information” than any of the corporate’s earlier fashions. Gemini 2.0 Pro can name instruments like Google Search, and execute code on behalf of customers.
Gemini 2.0 Pro’s context window is 2 million tokens, that means it might probably course of about 1.5 million phrases in a single go. Expressed one other method, Google’s latest AI mannequin might ingest all seven books within the Harry Potter sequence in a single immediate and nonetheless have about 400,000 phrases left over.
Google can be making its Gemini 2.0 Flash mannequin usually obtainable on Wednesday. This mannequin was introduced in December and is now obtainable to all customers of the Gemini app.
Lastly, presumably to rival the thrill surrounding DeepSeek’s fashions, Google is introducing a brand new, extra cost-efficient AI mannequin, Gemini 2.0 Flash-Lite. The firm says this mannequin outperforms its Gemini 1.5 Flash mannequin, however runs on the similar value and pace.
